A. Understanding the behavior of bioregulatory networks is critical for future progress in biology and therapeutics. Because of the complexity and high interconnectivity of these networks, reliable understanding can be achieved only with the aid of computer simulation studies. It is questionable however whether the present state of knowledge is sufficient for computational simulations of large systems to yield useful results. We took an alternative approach, based on the notion that essential behavior is already encoded in simple subsystems, and that further complexity serves to modulate this behavior. This possibility is attractive from considerations of evolution and system robustness. A second premise for the current investigation was that the response characteristics of the fundamental subsystems are often switch-like. We showed how this approach can facilitate linkage between theory and experiment. The subject of our investigation was the network that controls the induction of a set of genes in response to hypoxia. This was an attractive subject for theoretical study, because extensive (albeit yet incomplete) information has accumulated about relevant molecular and biological behavior, and because there is a clearly defined dependence of an output on an input. The input is the concentration of molecular oxygen, and the output is the activation of promoters that are under the control of hypoxia-regulated elements (HRE). We showed how a model ("core") subsystem can be selected from a molecular interaction map, how it can be encoded for computer survey of parameter space, how switch-like behavior can be found, and how the results may predict or guide experiments. This work was recently published (Kohn et al. 2004 Molec. Biol. Cell 15: 3042-52). We are collaborating with the Laboratory of Biosystems, CCR, NCI, which has been studying the hypoxia control system, we plan to test our theoretical predictions and to develop the theoretical model further guided by the results of experiments.B. In order to utilize most effectively the vast new information about the molecular interactions that make up the cell's regulatory networks, we urgently need a graphical method that will have the same utility that circuit diagrams serve for electronics. Complexity does not by itself make it hard to prepare diagrams. Metabolic pathway diagrams, for example, are commonly displayed on large wall charts. Bioregulatory networks, on the other hand, present special difficulties that are rarely encountered in the classical metabolic pathways: (1) function often depends on multimolecular complexes; (2) function is often modulated by multiple covalent modifications, such as phosphorylations of proteins; (3) bioregulatory proteins commonly consist of multiple domains with different functions, which may interact within the same molecule. We have developed a notation specifically suited for bioregulatory networks, and refer to the resulting diagrams as "molecular interaction maps" (because they can in fact be used like road maps to discern paths and connections) (Kohn 1999 Molec. Biol. Cell; Kohn 2001 Chaos; Aladjem et al. 2004 Science'). We have used this notation to generate maps of several key networks that function in cell proliferation control and have deposited some of them on John N. Weinstein's internet site: http://discover.nci.nih.gov. In collaboration with our LMP colleagues Mirit Aladjem and the Weinstein group, we have linked the maps with other bioinformatics databases. In addition, we are providing user-friendly features for navigation through the networks. Each molecular interaction is linked to an annotation that provides salient information and reference citations, so that the evidence bearing on each interaction can be evaluated.In addition to encyclopedic molecular interaction maps, which are analogous to road maps, we are preparing review articles that are analogous to guide-books. In these articles, we build a molecular interaction maps segmentally in a manner to illuminate functional organization, as well as provide comprehensive coverage of the known interactions. Currently in preparation are map-based reviews of the p53-Mdm2 system, the hypoxia control network, and the network centered on c-Myc.